graphid.core.mixin_loops module

class graphid.core.mixin_loops.InfrLoops

Bases: object

Algorithm control flow loops

main_gen(max_loops=None, use_refresh=True)

The main outer loop.

This function is designed as an iterator that will execute the graph algorithm main loop as automatically as possible, but if user input is needed, it will pause and yield the decision it needs help with. Once feedback is given for this item, you can continue the main loop by calling next. StopIteration is raised once the algorithm is complete.

Parameters:

• max_loops (int) – maximum number of times to run the outer loop, i.e. ranking is run at most this many times.

• use_refresh (bool) – allow the refresh criterion to stop the algo

Notes

Different phases of the main loop are implemented as subiterators

CommandLine

python -m graphid.core.mixin_loops InfrLoops.main_gen

Example

>>> from graphid.core.mixin_simulation import UserOracle
>>> from graphid import demo
>>> infr = demo.demodata_infr(num_pccs=3, size=5)
>>> infr.params['manual.n_peek'] = 10
>>> infr.params['ranking.ntop'] = 1
>>> infr.oracle = UserOracle(.99, rng=0)
>>> infr.simulation_mode = False
>>> infr.reset()
>>> gen = infr.main_gen()
>>> while True:
>>>     try:
>>>         reviews = next(gen)
>>>         edge, priority, data = reviews[0]
>>>         feedback = infr.request_oracle_review(edge)
>>>         infr.add_feedback(edge, **feedback)
>>>     except StopIteration:
>>>         break

hardcase_review_gen()

Subiterator for hardcase review

Re-review non-confident edges that vsone did not classify correctly

ranked_list_gen(use_refresh=True)

Subiterator for phase1 of the main algorithm

Calls the underlying ranking algorithm and prioritizes the results

incon_recovery_gen()

Subiterator for recovery mode of the mainm algorithm

Iterates until the graph is consistent

Note

inconsistency recovery is implicitly handled by the main algorithm, so other phases do not need to call this explicitly. This exists for the case where the only mode we wish to run is inconsistency recovery.

pos_redun_gen()

Subiterator for phase2 of the main algorithm.

Searches for decisions that would commplete positive redundancy

CommandLine

python -m graphid.core.mixin_loops InfrLoops.pos_redun_gen

Example

>>> from graphid.core.mixin_loops import *
>>> from graphid import demo
>>> infr = demo.demodata_infr(num_pccs=3, size=5, pos_redun=1)
>>> gen = infr.pos_redun_gen()
>>> feedback = next(gen)
>>> edge_ = feedback[0][0]
>>> print(edge_)
(1, 5)

neg_redun_gen()

Subiterator for phase3 of the main algorithm.

Searches for decisions that would commplete negative redundancy

_inner_priority_gen(use_refresh=False, only_auto=False)

Helper function that implements the general inner priority loop.

Executes reviews until the queue is empty or needs refresh

Parameters:

• user_refresh (bool) – if True enables the refresh criteria. (set to True in Phase 1)

• only_auto (bool) – reviews unless the graph is inconsistent. (set to True in Phase 3)

Notes

The caller is responsible for populating the priority queue. This will iterate until the queue is empty or the refresh critieron is triggered.

init_refresh()

start_id_review(max_loops=None, use_refresh=None)

main_loop(max_loops=None, use_refresh=True)

DEPRICATED

use list(infr.main_gen) instead or assert not any(infr.main_gen()) maybe this is fine.

class graphid.core.mixin_loops.InfrReviewers

Bases: object

try_auto_review(edge)

request_oracle_review(edge, **kw)

_make_review_tuple(edge, priority=None)

Makes tuple to be sent back to the user

emit_manual_review(edge, priority=None)

Emits a signal containing edges that need review. The callback should present them to a user, get feedback, and then call on_accpet.

skip(edge)

accept(feedback)

Called when user has completed feedback from qt or web

continue_review()